I'm small fry - nowhere near ST 2110 budgets - but I am relieved to hear so many of my own thoughts reflected from Doug. As a low-end production team with just a few members, 2110 seems like a "revisit in 2030" proposition. And honestly, I would be very dubious of investing in network or network-enabled equipment from BMD. They have a reputation for making versatile, budget-friendly equipment - but it's a given that QA isn't great, reliability isn't great, support isn't great and they'd rather put out a "new and improved" product instead of upgrading their existing ones. (Just look at Doug's Web Presenter / RUclips streaming woes)
it depends. 2110 Can help create realy simple setups, at least easy cabeling. if you use a bunch of the new studio cams and want to use them over a single network cable, the Blackmagic Studio Converter is about 1000$. A 24Port 10gbe poe Switch with 2x 100gb/s uplink port is just 3700$. if you se this with a atem that support 2110 ist getting cheaper than traditional sdi very fast.
@@juri14111996 No doubt the price will come down significantly. But if I'm spending >$4700 to convert a wrist-thick run of cables into a single cable, I want to be damn sure that one cable works. And reliability just isn't BMD's strong suit. Which is why their jump to 2110 is weird to me. It's like Wish deciding to sell parachutes.
As you note, uncompressed and compressed workflows have their benefits in overall topology cost and where they are used in the chain. A couple of notes on NDI. NDI discovery server (similar conceptually to NMOS) allows NDI to be registered and discovered across subnets, and also allows for practical uses of NDI for much larger device pools by speeding up the discovery versus the default automatic discovery implementation of NDI. Also the combination of various flavors of compression and NDI Bridge toolset allows the interconnect of NDI networks across the public internet. Unlike other protocols in that space, it does simplify one to many transmission when doing so. Thanks again for this video. Learned a lot!
Doug, Thank you for the explanation of SMPTE ST2110. I saw the video of NAB , but it was over my head, somewhat. You were very thorough; I will continue to read and try to stay abreast of it. Take care!
Great, relevance prioritized info as usual! It's great to see open, professional standards for IP network based A/V distribution! Since I only have interest in working with smaller scale production environments I had no idea that open standards like ST2110 were being implemented by the SMPTE organization, but after this video I feel I have a pretty solid general understanding of what they are and how they work.
This sounds like slightly redefined usage of the multicast ranges, the switches support pruning unless required instead of blast to determine required and then trim. Multicast itself has been in use for over a decade so it's a reliable system, still learning about precision time and this standards methods of integration on the network. This video is a nice step in that direction.
Great video and a good intro to 2110, thank you! (One minor point, "Vizrt" is usually pronounced Viz-r-t - i.e. the r and t are just pronounced as separate letters, so it sounds like: viz are tee)
Until BMD releases a 2110 switcher, their ecosystem is going to remain hybrid with all of its shortcomings. I would be surprised if we don't see a 2110 switcher out by next year, once they've got a little more experience with the ancillary stuff first. I think they'll sell a bunch of their little converters to pro markets for things like confidence monitors or non-critical interface conversion for things where a little latency isn't an issue. Random POV cameras, etc... BiDi over Cat6 worth PoE is very handy. I do think they'll sell a lot of the Audio monitors. That's right in the sweet spot for non-critical monitoring positions where a failure isn't a showstopper, and buying a dozen of them for a facility starts to have savings on any budget that can be then better spent elsewhere. Beyond that, the currently available 2110 switchers are in a very different class to most BMD gear. Which is a major differentiator in markets. BUT. "BMD 2110" is a talking point that shouldn't be ignored either.
From my experience with 2110 systems at major broadcasters in europe I dont see any of their 2110 products being used in professional systems. They are based on 10 GbE for endpoints, every professional system I've seen in the last years has been based on 25 GbE for (low channel count) endpoints with backbones at 100 GbE and up. This is due to the 12G bandwidth needed for 2160p50. Even though most broadcasters are still working in HD, nobody will build a new IP-system from scratch that doesn't offer a clear path to UHD. Blackmagic uses compression to do UHD over 10 GbE with their own codec, I don't see anybody else adopting that, as everybody else who wants to do compression uses JPEG XS. CAT wiring, which Blackmagic seems to push, although they do have some devices with fiber, is also not used for video networks. Many broadcast facilities also already have stuctured OS2-fiber cabling from their legacy SDI systems, so cost isn't that much of an issue. Another problem is that Blackmagic doesn't do 2022-7, which is the standard for dual network interfaces for redundancy, which every system I have seen uses. The lack of out-of-band-management is also a problem, in-band-managment isn't used much in my experience and only used in special circumstances (e.g. Grass Valley cameras due to a limitation of fibers in SMPTE 311 cables). Control is also a major issue in every system, the industry has embraces NMOS and Blackmagic say they support that, but in the real world it's often not as simple as that and interoperability is still a problem. And last but not least 2110 deployments require a lot more support by the manufacturer than SDI and in my experince with Blackmagic so far their support is pretty barebones. So I don't see how Blackmagic would fit into a professional system and the cost of integration would quickly make up the cost savings the hardware offers.
Great video, thanks for that. I'm always looking for alternative transmission methods in our old building with its thick walls. We only have 100Mbit from the switch in some areas and rarely more than 1Gbit between switches. That's why this is unfortunately not an option for us.
ndi cant realy be used if latency is importent. Thas the case is you need to show the live image on a projector or LED wall in the same space as the act is going on. NDI is great for livestreams and general improvicing.
That all sounds very interesting. Once word I don't think I heard was "security" which in most IP environments is crucial. Of course, once someone has physical access to your cables all bets are off anyway.
@@juri14111996 This is misleading. Security exists, but it's handled at a different level in the network architecture. 2110 is not really designed currently for low-end production systems.
It is clear that SMPTE is scoping this standard as a solution to high level productions. It solves the needs of big professionals and then it will slowly cascade to smaller productions and be democratized. I mean look at those big productions trucks with hundreds of SDI and big routers, this needs to evolve. Same as DANTE that is now the standard in big music events. SMPTE 2110 will take over definitely to traditional SDI wiring.
Fantastic video Doug! Fascinating stuff. Mikrotik makes a 24 sfp+ port 4 40gb switch that should be ptp aware for about $600, and those will start hitting the secondary market eventually. Beyond that, does the industry take the Dante approach and slap an FPGA to network card to make the whole system work at lowest possible latency? If we're using ethernet for AI applications surely it can be made fast enough for video transmission. It will be interesting to see. Low latency is paramount for me even above resolution so I'm sticking with coax and fiber SDI for the time being as well.
Hi Doug. We just got the 3x3 boxes. We've gotten them to work using direct connections. Our goal is to use it for basic video village setups. But what do I need to multicast to multiple gateways? The BM 360 switch isnt available yet. Any recos for a basic (cheap!) 10g switch that can allow multicast?
The bare minimum is a managed switch which supports IGMP Snooping v3. You'll need to enable that feature in the configuration interface (web or command-line). Once enabled globally, you'll need to enable it for each VLAN on your switch (if you have any setup), and then you may need to enable it for applicable each port on the switch as well. You'll also want to set options to tell your switch to discard unknown multicast traffic (otherwise it gets forwarded to all ports, flooding the network). I'd also enable the IGMP Snooping Querier if you can. You can enable Fast Leave if you want, but it isn't required. Consult the documentation for your switch or an experienced network professional to see how to do this, as it varies from model to model. Ideally you'd also want PTPv2 support, but finding that on "affordable" networking equipment is nigh unto impossible. The least expensive switch that I would trust that can do this is the TP-Link TL-SX3008F (amzn.to/463awxd) with compatible SFP+ modules like the TL-SM5310-T (amzn.to/4d3S1eq). But if you're willing to go with fiber connections rather than Ethernet you can actually save quite a bit, and use much longer cables. For that you'd use a single mode 10G SFP+ module (like amzn.to/3zG7zXc) on each end and a single mode cable, preferably armoured (like amzn.to/3W3L5ae). For this application fiber is quite a bit cheaper (and reliable at a distance) than copper. There are switches from other manufacturers that are designed and even pre-configured for AV over IP, starting with the Netgear M4300-8X8F. But the price does jump quite significantly.
Would love to learn more on how Ndi vs 2110 compares to each other. 😀👌 2110 or NDI video signals captured from within Resolume and Vmix for live stage presentations on huge led video walls on stage. 😀👌 Wow didn’t know there was 2110-10, 2110-20, 2110-30 and even 2110-21 😮 thank you for making these types of videos, more power to DJP 🎉😊
for 2110 with bmd you need an PCIE card, like decklink just with an sfp+ interface instead of BNC. Your standart Network Card is way to slow any way. NDI is completly different. is uses existing, low bandwith networks, so it needs to compress signal a lot and thus adds way more latency.
Is the blackmagic 2110 implementation/codec compatible with other vendors? For example, would the Panasonic AW-UE160 with 2110 work with the Blackmagic 2110 inputs?
great and detailed video as always! thanks. One big question, i do CCTV/IMAG work, normally 3 cameras on stage, connected (as now on 3 sdi cables) 75/100 meters to the FOH, we have a video switcher (atem extreme or studio) and vj on the FOH, and then the signal goes to the led screen on the stage... my question is... LATENCY? can i put one of the new BM 2110 3x3g boxes on stage and another 3x3g box at FOH, with one CAT cable, and get the same Latency as with SDI? do i need a PTP clock when only get box to box without other network? THANKS!!
The BMD 3x3G boxes have their own PTP time server that you can use, so that isn't an issue. But as far as latency, the best any of us have seen with those boxes is 2.5-3 frames.
@@djp_video thanks! i think 2-3 frames is a lot for someone singing live, in close frame, add the camera , and led screen processors delay and we are far away from audio lip sync
I have the video about Dante already. I haven't done one on NDI. I haven't spent enough time with NDI to feel confident enough in creating a video about it. (People creating videos about things they don't really know/understand is a pet peeve of mine.) But in any case, if you can put all of these technologies on their own independent network you'll be avoiding a lot of potential headaches. While you can run all of these on a single network, it's really not a great idea, and there's usually little reason to, as in many cases they don't interoperate with one another anyway and can start mess with each other's bandwidth or signaling or timing or whatever.
@@djp_video I agree completely! NDI should be distinct from the Dante network. I'm hopeful that 2110 would provide a superior alternative in combination with Dante
@@djp_video One of the main benefits of multiaccess networking is multiplexing many applications on shared networking equipment. The value of these technologies goes down significantly if they can not not share the same physical network.
Be aware, Ubiquiti switches do not handle IGMP v2 or v3 well. I know from experience. Also check that the backplane bandwidth supports all ports at full usage. Example 8 port 10gbps switch needs to handle minimum 80gbps throughput.
I think we need to demystify some assumptions. "2110 works with off-the-shelve equipment": no, it uses top-of-the-range switches that you can't find on Amazon. We're talking Arista in large installations.
The requirements aren't that stringent. IGMPv3 is found on most managed switches these days. And the PTP-aware switch doesn't seem to be strictly required for things to work. I did some of my testing on a TP-Link $250 switch and once I had IGMP Snooping setup correctly it all worked just fine. I wouldn't suggest it to someone, but it did work.
@@djp_video At small scale maybe, at a large scale I know a local broadcaster that has just spend half a milion on Arista switches for their 2110 facility. It is insane.
You could certainly spend a lot of money if you wanted to. I’ve got a small facility that only has 20 feeds, and we’ve had a lot of success with the Netgear Pro AV switches. They are 10G SFP and are a fraction of the cost of the Arista. They are PTP transparent and handle a large amount of Dante and 2110 traffic. However, you still need a controller. BMD’s 2110 switch is around $3k and has the controller built in. Perfect for trailers, conferences and small sporting events (I do a lot of high school events with 7-10 cameras. You can also get creative with multi-fiber runs. But if you’ve got the budget for Arista switches, you’re probably not looking at use-cases for the lower cost productions.
I'm waiting for the 2110 Codec to be released for Adobe so I can convert all my mp4 and HVEC vacation clips into to 2110 format so I can get the Hollywood look.
Crazy question for you. Thinking about the production trailer workflows, I wonder if an intermediate solution would be something similar to digital stageboxes for 2110? Cameras remain SDI, but all funnel down to a stagebox that translates it all into a trunk line, but because it is one box, could mitigate all of the timing issues collectively, rather than through a million converters, all vying to get that timing information on their own? Ideally, there could be a huge benefit where venues could have those boxes and all you would have to do is have a standardized trunk connection. Maybe utopian, but I look at how digital stageboxes have made impossibly complex setups not just workable, but far more capable than most of us ever imagined.
The issue is still the bandwidth as he mentioned 10gbit means 3 channels up and 3 channels down … if you get bigger connections you can do more trunking … very unlikely that venues give you this infrastructure outside having dark fiber
@@dv1858 I think it depends upon the venue, but with the cost of fiber a fraction of what it once was, it is a great way for newer venues to build out for the future.
@@Jason-mk3nn correct good fiber installation would be a nobrainer. At least in my region they still turning ever penny twice so I run cables all the time ;-)
From what I heared so far, you do not earlly want to connect your st2110 network to another one... just to mutch stuff that can break with the network... So even when to IP based Trucks are standing netxt to each other they will transmit signals over copper SDI Kabels... Having a good fiber setup in a location would be more benificial for setting up your own Stageboxes that are implementet in your network. and as fiber is relativly cheap thats something every location could easily provide in the location with ports outside on the parkinglot
@@jokiejo its the same with dante in Audio. if some others need audio connection, like the video/streaming devision you use xlr, even if both devision use dante internaly. alternativly they may use AES (2 channel Digital Audio over XLR) or a dante Gateway ( basical a Stagebox with just 2 sets of dante Cards, linked together internaly)
So I know the focus here is really about video productions and broadcasting but I would be interested in a solution there ST 2110 was implemented in a remote desktop solution. As you said, since there are no steps for encoding and decoding a video signal a pipeline properly implementing 2110 should have next to no latency. I could have my main computer in a closet and then simply remote into it at native speeds. It also sounds like you could support very odd display resolutions like any that follow the aspect ratios of 3:2, 16:10, or even like a 21:9. It sounds like the current implementations would not be that if there are sync issues but I can see a future where 10 or even 100Gbps routing is common and you simply connect to your server in this way.
It would likely require native support inside of a video card to "encode" the desktop into 2110-compatible data for good performance. While it isn't compressed, 2110 uses YCbCr (sometimes erroneously referred to as YUV) rather than RGB, and the conversion back and forth isn't necessarily complicated, it doesn't come for free. 1920x1080 at 60 fps is 124,416,000 pixels that have to be processed every second. Without hardware acceleration, that's potentially pretty processing intensive. Much of the live streaming of a desktop (Twich, Zoom, OBS, etc.) is accomplished using one of two methods: (1) the video card itself provides and/or encodes the data (since it has direct, raw access), or (2) a driver intercepts video data as the desktop is being drawn and builds an image from that, so it is effectively only having to process things that are changing rather than every pixel in every frame. Neither of those method gives full, direct access to uncompressed YCbCr data, so something has to do that conversion.
@@djp_video Yes the GPU would likely have to do the majority of the work and more than that there would probably need to be a gpu on the other side of the pipe to do the same. It is possible though. I think it is what intel is doing with their Intel Media Transport Library. I am pretty sure it is still in development as I am not aware of any products but that seems to be what they are doing.
@@Douglas_Gillette Ok this looks awesome. This is where I would like everything to go. IPMX looks like a great standard, It is an open standard, It also has mechanisms to apply optional codecs depending on the bandwidth available. If someone has 100Gbps networking then sure they can go with an uncompressed stream but in the more likely scenario that someone has 10Gbps or less then they can still use the implementation with only slightly higher latency.
2110 can be implemented with low latency. In the case of Blackmagic they've designed their converter products with buffering on both ends. It doesn't have to be that way.
I'm small fry - nowhere near ST 2110 budgets - but I am relieved to hear so many of my own thoughts reflected from Doug. As a low-end production team with just a few members, 2110 seems like a "revisit in 2030" proposition.
And honestly, I would be very dubious of investing in network or network-enabled equipment from BMD. They have a reputation for making versatile, budget-friendly equipment - but it's a given that QA isn't great, reliability isn't great, support isn't great and they'd rather put out a "new and improved" product instead of upgrading their existing ones. (Just look at Doug's Web Presenter / RUclips streaming woes)
it depends. 2110 Can help create realy simple setups, at least easy cabeling.
if you use a bunch of the new studio cams and want to use them over a single network cable, the Blackmagic Studio Converter is about 1000$.
A 24Port 10gbe poe Switch with 2x 100gb/s uplink port is just 3700$.
if you se this with a atem that support 2110 ist getting cheaper than traditional sdi very fast.
@@juri14111996 No doubt the price will come down significantly. But if I'm spending >$4700 to convert a wrist-thick run of cables into a single cable, I want to be damn sure that one cable works. And reliability just isn't BMD's strong suit. Which is why their jump to 2110 is weird to me. It's like Wish deciding to sell parachutes.
As you note, uncompressed and compressed workflows have their benefits in overall topology cost and where they are used in the chain. A couple of notes on NDI.
NDI discovery server (similar conceptually to NMOS) allows NDI to be registered and discovered across subnets, and also allows for practical uses of NDI for much larger device pools by speeding up the discovery versus the default automatic discovery implementation of NDI. Also the combination of various flavors of compression and NDI Bridge toolset allows the interconnect of NDI networks across the public internet. Unlike other protocols in that space, it does simplify one to many transmission when doing so. Thanks again for this video. Learned a lot!
Doug,
Thank you for the explanation of SMPTE ST2110. I saw the video of NAB , but it was over my head, somewhat. You were very thorough; I will continue to read and try to stay abreast of it. Take care!
Thanks Doug, great video. Loved the honest comments and your feedback on 2110.
Thanks again Doug for another great video
Great, relevance prioritized info as usual! It's great to see open, professional standards for IP network based A/V distribution! Since I only have interest in working with smaller scale production environments I had no idea that open standards like ST2110 were being implemented by the SMPTE organization, but after this video I feel I have a pretty solid general understanding of what they are and how they work.
This sounds like slightly redefined usage of the multicast ranges, the switches support pruning unless required instead of blast to determine required and then trim. Multicast itself has been in use for over a decade so it's a reliable system, still learning about precision time and this standards methods of integration on the network. This video is a nice step in that direction.
Great video and a good intro to 2110, thank you! (One minor point, "Vizrt" is usually pronounced Viz-r-t - i.e. the r and t are just pronounced as separate letters, so it sounds like: viz are tee)
I've heard both, even from intdustry veterans. But yes, your suggested pronunciation is more common.
Until BMD releases a 2110 switcher, their ecosystem is going to remain hybrid with all of its shortcomings. I would be surprised if we don't see a 2110 switcher out by next year, once they've got a little more experience with the ancillary stuff first.
I think they'll sell a bunch of their little converters to pro markets for things like confidence monitors or non-critical interface conversion for things where a little latency isn't an issue. Random POV cameras, etc... BiDi over Cat6 worth PoE is very handy.
I do think they'll sell a lot of the Audio monitors. That's right in the sweet spot for non-critical monitoring positions where a failure isn't a showstopper, and buying a dozen of them for a facility starts to have savings on any budget that can be then better spent elsewhere.
Beyond that, the currently available 2110 switchers are in a very different class to most BMD gear. Which is a major differentiator in markets.
BUT. "BMD 2110" is a talking point that shouldn't be ignored either.
From my experience with 2110 systems at major broadcasters in europe I dont see any of their 2110 products being used in professional systems.
They are based on 10 GbE for endpoints, every professional system I've seen in the last years has been based on 25 GbE for (low channel count) endpoints with backbones at 100 GbE and up. This is due to the 12G bandwidth needed for 2160p50. Even though most broadcasters are still working in HD, nobody will build a new IP-system from scratch that doesn't offer a clear path to UHD. Blackmagic uses compression to do UHD over 10 GbE with their own codec, I don't see anybody else adopting that, as everybody else who wants to do compression uses JPEG XS.
CAT wiring, which Blackmagic seems to push, although they do have some devices with fiber, is also not used for video networks. Many broadcast facilities also already have stuctured OS2-fiber cabling from their legacy SDI systems, so cost isn't that much of an issue.
Another problem is that Blackmagic doesn't do 2022-7, which is the standard for dual network interfaces for redundancy, which every system I have seen uses.
The lack of out-of-band-management is also a problem, in-band-managment isn't used much in my experience and only used in special circumstances (e.g. Grass Valley cameras due to a limitation of fibers in SMPTE 311 cables).
Control is also a major issue in every system, the industry has embraces NMOS and Blackmagic say they support that, but in the real world it's often not as simple as that and interoperability is still a problem.
And last but not least 2110 deployments require a lot more support by the manufacturer than SDI and in my experince with Blackmagic so far their support is pretty barebones.
So I don't see how Blackmagic would fit into a professional system and the cost of integration would quickly make up the cost savings the hardware offers.
@janludwig7286 I don't think Blackmagic is targeting that type of customer at all. They're clearly after much smaller operations.
Great video, thanks for that. I'm always looking for alternative transmission methods in our old building with its thick walls. We only have 100Mbit from the switch in some areas and rarely more than 1Gbit between switches. That's why this is unfortunately not an option for us.
I assume that means that running fiber isn't an option? Because that would give you all the bandwidth you could ever need.
@@djp_video Hi Doug, that's exactly what we are doing with dedicated connections. Our data network doesn't have enough bandwidth
Thanks for explaining this to us, Doug! Great video!
Thanks for explaining all this. It’s confusing after hearing a lot of hype about NDI, but I think I get it now.
ndi cant realy be used if latency is importent. Thas the case is you need to show the live image on a projector or LED wall in the same space as the act is going on.
NDI is great for livestreams and general improvicing.
That all sounds very interesting. Once word I don't think I heard was "security" which in most IP environments is crucial. Of course, once someone has physical access to your cables all bets are off anyway.
security in av networking is not a thing. most devices dont have any login at all (BMD ATEM, Yamaha Audio Mixers, ......)
@@juri14111996 This is misleading. Security exists, but it's handled at a different level in the network architecture. 2110 is not really designed currently for low-end production systems.
What are some threats targeted at AV systems? I know the Department of Defense will care about this.
Thanks Doug 😊😊😊
It is clear that SMPTE is scoping this standard as a solution to high level productions. It solves the needs of big professionals and then it will slowly cascade to smaller productions and be democratized. I mean look at those big productions trucks with hundreds of SDI and big routers, this needs to evolve. Same as DANTE that is now the standard in big music events. SMPTE 2110 will take over definitely to traditional SDI wiring.
Fantastic video Doug! Fascinating stuff. Mikrotik makes a 24 sfp+ port 4 40gb switch that should be ptp aware for about $600, and those will start hitting the secondary market eventually. Beyond that, does the industry take the Dante approach and slap an FPGA to network card to make the whole system work at lowest possible latency? If we're using ethernet for AI applications surely it can be made fast enough for video transmission. It will be interesting to see. Low latency is paramount for me even above resolution so I'm sticking with coax and fiber SDI for the time being as well.
Your videos are always informative and usefull. Thank you for your efforts!❤
Hi Doug. We just got the 3x3 boxes. We've gotten them to work using direct connections. Our goal is to use it for basic video village setups. But what do I need to multicast to multiple gateways? The BM 360 switch isnt available yet. Any recos for a basic (cheap!) 10g switch that can allow multicast?
The bare minimum is a managed switch which supports IGMP Snooping v3. You'll need to enable that feature in the configuration interface (web or command-line). Once enabled globally, you'll need to enable it for each VLAN on your switch (if you have any setup), and then you may need to enable it for applicable each port on the switch as well. You'll also want to set options to tell your switch to discard unknown multicast traffic (otherwise it gets forwarded to all ports, flooding the network). I'd also enable the IGMP Snooping Querier if you can. You can enable Fast Leave if you want, but it isn't required.
Consult the documentation for your switch or an experienced network professional to see how to do this, as it varies from model to model.
Ideally you'd also want PTPv2 support, but finding that on "affordable" networking equipment is nigh unto impossible.
The least expensive switch that I would trust that can do this is the TP-Link TL-SX3008F (amzn.to/463awxd) with compatible SFP+ modules like the TL-SM5310-T (amzn.to/4d3S1eq). But if you're willing to go with fiber connections rather than Ethernet you can actually save quite a bit, and use much longer cables. For that you'd use a single mode 10G SFP+ module (like amzn.to/3zG7zXc) on each end and a single mode cable, preferably armoured (like amzn.to/3W3L5ae). For this application fiber is quite a bit cheaper (and reliable at a distance) than copper.
There are switches from other manufacturers that are designed and even pre-configured for AV over IP, starting with the Netgear M4300-8X8F. But the price does jump quite significantly.
All that aes67 goodness as well. Yummy stuff
Would love to learn more on how Ndi vs 2110 compares to each other. 😀👌
2110 or NDI video signals captured from within Resolume and Vmix for live stage presentations on huge led video walls on stage. 😀👌
Wow didn’t know there was
2110-10, 2110-20, 2110-30 and even 2110-21 😮 thank you for making these types of videos, more power to DJP 🎉😊
for 2110 with bmd you need an PCIE card, like decklink just with an sfp+ interface instead of BNC.
Your standart Network Card is way to slow any way.
NDI is completly different. is uses existing, low bandwith networks, so it needs to compress signal a lot and thus adds way more latency.
Is the blackmagic 2110 implementation/codec compatible with other vendors? For example, would the Panasonic AW-UE160 with 2110 work with the Blackmagic 2110 inputs?
It's supposed to be. At least for anything lower than 2160p30. (Camera limitation, not BMD.) The only way to know for sure is to try it.
great and detailed video as always! thanks. One big question, i do CCTV/IMAG work, normally 3 cameras on stage, connected (as now on 3 sdi cables) 75/100 meters to the FOH, we have a video switcher (atem extreme or studio) and vj on the FOH, and then the signal goes to the led screen on the stage... my question is... LATENCY? can i put one of the new BM 2110 3x3g boxes on stage and another 3x3g box at FOH, with one CAT cable, and get the same Latency as with SDI? do i need a PTP clock when only get box to box without other network? THANKS!!
The BMD 3x3G boxes have their own PTP time server that you can use, so that isn't an issue. But as far as latency, the best any of us have seen with those boxes is 2.5-3 frames.
@@djp_video thanks! i think 2-3 frames is a lot for someone singing live, in close frame, add the camera , and led screen processors delay and we are far away from audio lip sync
I've been waiting for this video for a long time! YAY Deej
The Blackmagic 2110 IP Converter 3x3G is on my wishlist lol
As I’ve studied this seems that my next installation will ne designed around 2110 Just due simplicity of cabling.
Thank you! You should make a video about NDI and Dante. NDI and Dante on a same network is a nightmare. Is 2110 going to solve that problem?
I have the video about Dante already. I haven't done one on NDI. I haven't spent enough time with NDI to feel confident enough in creating a video about it. (People creating videos about things they don't really know/understand is a pet peeve of mine.)
But in any case, if you can put all of these technologies on their own independent network you'll be avoiding a lot of potential headaches. While you can run all of these on a single network, it's really not a great idea, and there's usually little reason to, as in many cases they don't interoperate with one another anyway and can start mess with each other's bandwidth or signaling or timing or whatever.
@@djp_video I agree completely! NDI should be distinct from the Dante network. I'm hopeful that 2110 would provide a superior alternative in combination with Dante
@@djp_video One of the main benefits of multiaccess networking is multiplexing many applications on shared networking equipment. The value of these technologies goes down significantly if they can not not share the same physical network.
Be aware, Ubiquiti switches do not handle IGMP v2 or v3 well. I know from experience. Also check that the backplane bandwidth supports all ports at full usage. Example 8 port 10gbps switch needs to handle minimum 80gbps throughput.
160 Gbps for full duplex. I learned that this is called non-blocking.
I think we need to demystify some assumptions. "2110 works with off-the-shelve equipment": no, it uses top-of-the-range switches that you can't find on Amazon. We're talking Arista in large installations.
same with AVB in audio, but AVB is fine with just 1gbe.
some mikrotik router/Switch might work. an theyr "cheap".
The requirements aren't that stringent. IGMPv3 is found on most managed switches these days. And the PTP-aware switch doesn't seem to be strictly required for things to work. I did some of my testing on a TP-Link $250 switch and once I had IGMP Snooping setup correctly it all worked just fine. I wouldn't suggest it to someone, but it did work.
@@djp_video At small scale maybe, at a large scale I know a local broadcaster that has just spend half a milion on Arista switches for their 2110 facility. It is insane.
You could certainly spend a lot of money if you wanted to. I’ve got a small facility that only has 20 feeds, and we’ve had a lot of success with the Netgear Pro AV switches. They are 10G SFP and are a fraction of the cost of the Arista. They are PTP transparent and handle a large amount of Dante and 2110 traffic. However, you still need a controller. BMD’s 2110 switch is around $3k and has the controller built in. Perfect for trailers, conferences and small sporting events (I do a lot of high school events with 7-10 cameras. You can also get creative with multi-fiber runs. But if you’ve got the budget for Arista switches, you’re probably not looking at use-cases for the lower cost productions.
I'm waiting for the 2110 Codec to be released for Adobe so I can convert all my mp4 and HVEC vacation clips into to 2110 format so I can get the Hollywood look.
Don't forget the bokeh and LUT plugins too
Crazy question for you. Thinking about the production trailer workflows, I wonder if an intermediate solution would be something similar to digital stageboxes for 2110? Cameras remain SDI, but all funnel down to a stagebox that translates it all into a trunk line, but because it is one box, could mitigate all of the timing issues collectively, rather than through a million converters, all vying to get that timing information on their own? Ideally, there could be a huge benefit where venues could have those boxes and all you would have to do is have a standardized trunk connection. Maybe utopian, but I look at how digital stageboxes have made impossibly complex setups not just workable, but far more capable than most of us ever imagined.
The issue is still the bandwidth as he mentioned 10gbit means 3 channels up and 3 channels down … if you get bigger connections you can do more trunking … very unlikely that venues give you this infrastructure outside having dark fiber
@@dv1858 I think it depends upon the venue, but with the cost of fiber a fraction of what it once was, it is a great way for newer venues to build out for the future.
@@Jason-mk3nn correct good fiber installation would be a nobrainer.
At least in my region they still turning ever penny twice so I run cables all the time ;-)
From what I heared so far, you do not earlly want to connect your st2110 network to another one... just to mutch stuff that can break with the network... So even when to IP based Trucks are standing netxt to each other they will transmit signals over copper SDI Kabels...
Having a good fiber setup in a location would be more benificial for setting up your own Stageboxes that are implementet in your network. and as fiber is relativly cheap thats something every location could easily provide in the location with ports outside on the parkinglot
@@jokiejo its the same with dante in Audio. if some others need audio connection, like the video/streaming devision you use xlr, even if both devision use dante internaly.
alternativly they may use AES (2 channel Digital Audio over XLR) or a dante Gateway ( basical a Stagebox with just 2 sets of dante Cards, linked together internaly)
Its going to be when the ATEM vision mixers can take SMPTE 2110. I think it will be a ATEM with 100gbe port
So I know the focus here is really about video productions and broadcasting but I would be interested in a solution there ST 2110 was implemented in a remote desktop solution. As you said, since there are no steps for encoding and decoding a video signal a pipeline properly implementing 2110 should have next to no latency. I could have my main computer in a closet and then simply remote into it at native speeds. It also sounds like you could support very odd display resolutions like any that follow the aspect ratios of 3:2, 16:10, or even like a 21:9. It sounds like the current implementations would not be that if there are sync issues but I can see a future where 10 or even 100Gbps routing is common and you simply connect to your server in this way.
It would likely require native support inside of a video card to "encode" the desktop into 2110-compatible data for good performance. While it isn't compressed, 2110 uses YCbCr (sometimes erroneously referred to as YUV) rather than RGB, and the conversion back and forth isn't necessarily complicated, it doesn't come for free. 1920x1080 at 60 fps is 124,416,000 pixels that have to be processed every second. Without hardware acceleration, that's potentially pretty processing intensive.
Much of the live streaming of a desktop (Twich, Zoom, OBS, etc.) is accomplished using one of two methods: (1) the video card itself provides and/or encodes the data (since it has direct, raw access), or (2) a driver intercepts video data as the desktop is being drawn and builds an image from that, so it is effectively only having to process things that are changing rather than every pixel in every frame. Neither of those method gives full, direct access to uncompressed YCbCr data, so something has to do that conversion.
@@djp_video Yes the GPU would likely have to do the majority of the work and more than that there would probably need to be a gpu on the other side of the pipe to do the same. It is possible though. I think it is what intel is doing with their Intel Media Transport Library. I am pretty sure it is still in development as I am not aware of any products but that seems to be what they are doing.
IPMX, the continuation of ST 2110, holds the solution to unique resolution formats.
@@Douglas_Gillette Ok this looks awesome. This is where I would like everything to go. IPMX looks like a great standard, It is an open standard, It also has mechanisms to apply optional codecs depending on the bandwidth available. If someone has 100Gbps networking then sure they can go with an uncompressed stream but in the more likely scenario that someone has 10Gbps or less then they can still use the implementation with only slightly higher latency.
does 2110 support PoE?
Devices which use ST 2110 can also support PoE. The two technologies are independent of one another.
33:44 the latency issue, yes indeedy. It's gonna have some frames of latency. No ifs, ands, or buts.
2110 can be implemented with low latency. In the case of Blackmagic they've designed their converter products with buffering on both ends. It doesn't have to be that way.
Why are they stuffing in these meaningless words to describe existing concepts? Essence? It's called SOURCE! Its like calling tasks stories. So dumb!
"oh no, I've lost my essence" can you imagine?!!